Generally, an apparatus and a method for encoding multiview video downscale two views of the multiview video, determine a shift between the downscaled views, and encode one the two views using the shift.
Legal claims defining the scope of protection, as filed with the USPTO.
1. A method for multiview video coding, the method comprising: downscaling two views of a multiview video; determining a shift between the two downscaled views; encoding one of the two views using the shift; upscaling the shift; identifying co-located areas of the two views using the upscaled shift; comparing pixels of the co-located areas in a first comparison, shifting one of the co-located areas by a predetermined shift amount; comparing pixels of the other of the co-located areas to pixels of the shifted co-located area in a second comparison; and refining the upscaled shift with the predetermined shift amount if the second comparison results in less residue than the first comparison.
2. A method as in claim 1 , wherein the two views include a base view and a dependent view of a multiview image, and wherein encoding one of the two views using the shift includes encoding the dependent view using the shift.
3. A method as in claim 1 , wherein the two views include a view from a multiview image and a view from a preceding multiview image.
4. A method as in claim 1 , further comprising receiving a mode command from a user and transmitting a mode signal based on the mode command to select between fast motion search logic and another motion search logic, the fast motion search logic configured to downscale the two views, determine the shift between the two downscaled views, and upscale the shift, wherein the fast motion search logic requires less power than the another motion search logic.
5. A method as in claim 1 , wherein the shift comprises an inter-view shift, further comprising downscaling a preceding view from a preceding multiview image, determining a predicted shift between the downscaled preceding view and one of the two downscaled views, and determining an optimal shift between the inter-view shift and the predicted shift, wherein encoding one of the two views using the shift comprises encoding one of the two views using the optimal shift.
6. A method as in claim 5 , wherein the preceding view is a preceding dependent view, the two views include a dependent view, and the predicted shift comprises a shift prediction between the preceding dependent view and the dependent view.
7. A method as in claim 1 , wherein the shift comprises an inter-view shift, further comprising identifying co-located areas of the two views using the inter-view shift, and identifying a corresponding co-located area of a downscaled preceding view of a preceding multiview image.
8. A method as in claim 7 , further comprising upscaling the inter-view shift, wherein using the inter-view shift comprises using the upscaled inter-view shift.
9. A method as in claim 7 , further comprising determining a predicted shift between another downscaled preceding view of the preceding multiview image and one of the two downscaled views; and using the predicted shift to identify the corresponding co-located area.
10. A method as in claim 7 , wherein identifying a corresponding co-located area comprises determining motion vectors of macroblocks; and averaging the macroblock motion vectors.
11. A method as in claim 7 , wherein each of the two views has a total area, and the co-located areas are smaller than the total area of each of the two views.
12. A method as in claim 1 , wherein comparing pixels comprises comparing only pixels located near corners of at least one of the co-located areas.
13. A method as in claim 1 , wherein the predetermined shift amount is less than a pixel.
14. A method as in claim 1 , wherein the shifted co-located area corresponds to a base view.
15. A method as in claim 1 , further comprising obtaining, with a multiview source device, the multiview video.
16. A method as in claim 15 , wherein the multiview device includes a display device and a user interface, further comprising receiving a mode command with the user interface and, responsive to the mode command, encoding the multiview video with one of fast motion search logic and another motion search logic, wherein the fast motion search logic downscales the two views, determines the shift between the two downscaled views, and encodes one of the tow views using the shift.
17. An apparatus for encoding multiview video, the apparatus comprising: fast motion search logic operable to downscale two views of the multiview video and determine a shift between the two downscaled views; encoder logic configured to encode one of the two views using the shift; a motion vector upscaler configured to generate an upscaled shift by upscaling the shift; and a motion vector refiner configured to refine the upscaled shift to improve the quality of the encoded view, wherein the motion vector refiner is configured to refine the upscaled shift by identifying co-located areas of the two views using the upscaled shift; comparing pixels of the co-located areas in a first comparison; shifting one of the co-located areas by a predetermined shift amount; comparing pixels of the other of the co-located areas to pixels of the shifted co-located area in a second comparison; and refining the upscaled shift with the predetermined shift amount if the second comparison results in less residue than the first comparison.
18. An apparatus as in claim 17 , further comprising an integrated circuit, wherein the fast motion search logic is included in the integrated circuit.
19. An apparatus as in claim 17 , further comprising a user interface configured to receive a mode command from a user and select, responsive to the mode command, between the fast motion search logic and another motion search logic, the fast motion search logic consuming less power than the another motion search logic.
20. An apparatus as in claim 17 , wherein the shift comprises an inter-view shift, and the fast motion search logic is configured to downscale a preceding view from a preceding multiview image, determine a predicted shift between the downscaled preceding view and one of the two downscaled views, and determine an optimal shift between the inter-view shift and the predicted shift, wherein using the shift comprises using the optimal shift.
21. An apparatus as in claim 17 , wherein the motion vector refiner is configured to compare only pixels located near corners of at least one of the co-located areas.
22. An apparatus as in claim 17 , wherein the predetermined shift amount is less than a pixel.
23. An apparatus as in claim 17 , further comprising a multiview source device operable to capture or generate the multiview video.
24. An apparatus as in claim 23 , wherein the multiview source device is selected from a group comprising a computing device, a video camera, an automobile safety camera, a security system, a tablet, a smart phone, a gaming console, a video conferencing system, a television set top box, and an apparatus capable of receiving or generating video.
25. An integrated circuit for use in encoding a multiview video, the integrated circuit comprising: fast motion search logic operable to downscale two views of the multiview video and determine a shift between the two downscaled views; encoder logic configured to encode one of the two views using the shift; a motion vector u scaler configured to generate an upscaled shift by upscaling the shift; and a motion vector refiner configured to refine the upscaled shift to improve the quality of the encoded view, wherein the motion vector refiner is configured to refine the upscaled shift by identifying co-located areas of the two views using the upscaled shift, comparing pixels of the co-located areas in a first comparison; shifting one of the co-located areas by a predetermined shift amount comparing pixels of the other of the co-located areas to pixels of the shifted co-located area in a second comparison; and refining the upscaled shift with the predetermined shift, amount if the second comparison results in less residue than the first comparison.
26. An integrated circuit as in claim 25 , further comprising an imaging sensor operable to capture the multiview video.
27. A non-transitory computer readable medium comprising executable instructions that when executed by an integrated circuit fabrication system, cause the integrated circuit fabrication system to produce an integrated circuit that comprises: fast motion search logic operable to downscale two views of the multiview video and determine a shift between the two downscaled views; encoder logic configured to encode one of the two views using the shift; a motion vector upscaler configured to generate an upscaled shift by upscaling the shift; and a motion vector refiner configured to refine the upscaled shift to improve the quality of the encoded view, wherein the motion vector refiner is configured to refine the upscaled shift by identifying co-located areas of the two views using the upscaled shift; comparing pixels of the co-located areas in a first comparison; shifting one of the co-located areas by a predetermined shift amount; comparing pixels of the other of the co-located areas to pixels of the shifted co-located area in a second comparison and refining the upscaled shift with the predetermined shift amount if the second comparison results in less residue than the first comparison.
Cooperative Patent Classification codes for this invention. Click any code to explore related patents in that topic.
July 15, 2013
November 15, 2016
Browse 5M+ US patents with plain-English claim translations and AI-generated analysis.